I am certainly not going to weigh in on which is more accurate but have these two models (Piomass and DMI) diverged to this extent in the past? If not, this is very likely due to the fact that some of the model assumptions are breaking down as a result of the entirely new regime in the cryosphere. This could be happening to either or perhaps both of the models.
Answer to your question was given by me in this very topic,
today. Given that 2012 was quite unusual year with tons of melt overall, your assumption sounds only more reasonable, to me.
And one more strange thing about this year's massive disagreement between PIOMASS and DMI. Previous four years, despite all bumps and weathers, i see very fine agreement between those systems in terms of "places" for June melt. I.e., the year with most June melt is the same year in both models, the year with least mealt - the same, and all in between. Except this year, that is. Look:
ASI June volume loss, x1000 km^3:
2011: PIOMASS 6,7; DMI 8,7
2012: PIOMASS 7,3; DMI 10,4
2013: PIOMASS 6,6; DMI 8,0
2014: PIOMASS 5,7; DMI 6,4
I.e., 2012 is most ice lost, 2nd-most is 2011, 3rd-most is 2013, and least ice lost is 2014 - all that in BOTH models. But now we get:
2015: PIOMASS 6,2; DMI 9,2
Which is 4th place out of five years in PIOMASS, but 2nd place in DMI. This is exactly what i was talking few pages ago: models can have all sorts of varying assumptions, but usually their behaviour over time about reflecting one and same physical process - correlates in many detectable ways. This year's June, it doesn't. Which is why i said someone went nuts in either PIOMASS or DMI. That's of course figurative. Perhaps data is damaged within one of them, or there is intended unfair play, etc.
I hope, Neven, you do respect my dropped pants a tiny bit more now.
There is logic in here observation, and i suspect you see it easily.